Saturday, January 08, 2011

'Ongoing Victim Suffering Increases Prejudice:The Case of Secondary Anti-Semitism'

'Ongoing Victim Suffering Increases Prejudice: The Case of Secondary Anti-Semitism,' a paper by Roland Imhoff and Rainer Banse, is a worthwhile read if you're interested in how one's empathy towards suffering doesn't particularly extend to suffering that was initiated in the distant past but is still ongoing. Here's the abstract:
Some people have postulated that the perception of Jews’ ongoing suffering from past atrocities can result in an increase in anti-Semitism. This postulated secondary anti-Semitism is compatible with a number of psychological theories, but until now there has been no empirical evidence in support of this notion. The present study provides the first evidence that ongoing suffering evokes an increase in prejudice against the victims. However, this effect became apparent only if respondents felt obliged to respond truthfully because of a bogus pipeline (BPL); without this constraint, the perception of ongoing victim suffering led to a socially desirable reduction in self reported prejudice. The validity of the BPL manipulation was confirmed by the finding that itmoderated the relation between explicit and implicit anti-Semitism, as measured with an affect misattribution procedure.
And here's the first paragraph:
‘The Germans will never forgive the Jews for Auschwitz.’’ This famous quip by the Israeli psychoanalyst Zvi Rex (e.g., Buruma, 2003) describes the core assumption of secondary-anti-Semitism theory, which suggests that every reminder of the German
atrocities and the victims’ suffering still evokes aversive feelings of guilt and thus increases a defensive anti-Semitism—even in Germans born decades after 1945 (e.g., Adorno, 1955; Bergmann, 2006). Despite the popularity of this assumption in the
social sciences and its compatibility with a number of psychological theories, it has never been empirically tested.
Because the paper is concise and quite strong, I don't have anything to say about the content, but I do have something to say about the context. In order to measure explicit anti-Semitism, the researchers used a 29-item agree-disagree survey. You can get the whole survey by e-mailing the authors but here's the only item that's cited in the paper: "Jews have too much influence on public opinion." The problem with this statement as a measure, which the authors neglect to discuss, is that in the current context, Israel is perceived as a Jewish nation, and Jewish people are semantically connected with Israel and the United States. (And within the U.S., Jewish people are--at least semantically--more connected with AIPAC than, say, J Street.) So it's impossible to tell whether people answering this item were really responding to "Jews have too much influence." or "Israel has too much influence" or "Israel and AIPAC have too much influence." I don't think any sort of rephrasing would help either. "Jews, apart from those in the U.S. and Israel, have too much influence on public opinion" might separate the U.S. and Israel, but it still makes them salient in a reader's mind.

'Modern Anti-Semitism and Anti-Israeli Attitudes' is another paper where this same contextual issue occurs. In this study, the authors (including Lee Jussim, who does some fascinating research) measure attitudes towards Israel, India and Russia to see if people are more severely critical of Israel's activities in Palestine, than they are towards India's activities in Kashmir and Russia's activities in Chechnya. The problem here is even more obvious. The authors are correctly relying on the notion that Israel equals Jewish people for most people, but then concluding that their findings apply to measures of anti-Semitism, when they actually apply to the conflated concept of anti-Israeliness, which includes some uncertain element of anti-Semitism.

The same problem would occur if you're measuring anti-Muslim attitudes either by (a) using Saudi Arabia as a proxy for Muslims or (b) neglecting to subtract any anti-Saudi element from your anti-Muslim variable. It would also occur if you're measuring anti-Hindu attitudes either by (a) using India as a proxy for Hindus or (b) neglected to subtract any anti-India attitudes from your anti-Hindu variable. Because some countries in the world are inextricably associated with certain religions, you'll probably end up measuring some conflated concept, which absolutely deserves attention but nonetheless deserves new nomenclature. The old nomenclature is too precise.